A Study On Chinese Calligraphy Beautification Approach From Handwritten Characters For Robotic Writing

Chinese calligraphy,which is an ancient art of Chinese characters,is one of the basic elements of Chinese culture.It is highly valued all over the world,especially in East Asia.Although calligraphy is popular with the masses,it is a challenge for non-expert users to write visually pleasing calligraphy with his or her own unique style.In recent years,the calligraphy beautification with a specific style for handwritten Chinese characters has become a research hotspot.However current researches focus on simulating digital calligraphy with certain style,instead of generating calligraphy characters that can be physically written on paper with a brush.In this paper,we introduce the convolutional neural network and optimization method to implement an intelligent calligraphy beautification and writing system based on robotic arm.First,we sketch the handwritten characters using a mouse or a touch pad.Then,the system will identify the stroke and achieve intelligent beautification in kai style.Next,the robotic arms are controlled to write calligraphy characters on the rice paper with a brush.The writing results show that our system can achieve visually pleasing calligraphy from various input handwriting while retaining the user's style.The main contents in this research are as follows:(1)Research on methods of interactive handwriting acquisition and stroke recognition.Aiming at the difficulty of stroke segmentation and recognition of handwritten Chinese characters based on image,we sketch the handwritten characters in the form of skeleton using a mouse or a touch pad.Skeleton points on each stroke are synchronously collected and saved in a file.The stroke extraction rate is almost one stroke per second,which is much higher than the efficiency of stroke extraction based on Chinese image.Then,skeleton points of each stroke automatically generate the stroke image without noise.The improved convolutional neural network(CNN)is used to identify individual stroke and retrieve the corresponding standard strokes.The average recognition accuracy reaches 93%.It can meet the requirements of follow-up research.(2)Research on Chinese calligraphy beautification approach from handwritten characters for robotic writing.It is difficult to output aesthetically beautiful calligraphy with the user's style when the handwritten Chinese characters are converted into the handwritten Chinese characters,so a standard calligraphy stroke library for robotic arm based on Tian Yingzhang's and Tian Xuesong's strokes is constructed.Then a shape character vector is presented to describe the shape of standard strokes,and a global optimization approach to solve the minimization problem between the handwritten strokesand standard calligraphy strokes is proposed.(3)Controlling methods for robotic arms to write calligraphy with a brush.There is no doubt that it is difficult to control the height and force of the brush during writing calligraphy.In the paper,the optimized stroke trajectory is input to the robotic arm to form a motion trajectory on the 2D plane.After selecting the height of the relocation points on the standard calligraphy stroke,the height value in 3D space will be added for every 2D control point.After that coordinate control method is used to make the robotic arm write calligraphy in Kai style.To achieve more realistic Kai style,the rigid part in every stroke is strengthen.Calligraphy written by the robotic arm are not only demonstrate full regular script style,but also show the characteristics of users' handwriting.(4)Evaluation method of the calligraphy results writing by the robotic arm.In order to solve the problem that there is no effective evaluation model for robotic arm,a calligraphy simulation method is presented,and the digital calligraphy is compared with real calligraphy.Results of similarity comparison in three aspects are respectively achieved.Among them,the structure ratio is above 92%,the unit ratio is above 90%,and the inclination degree is above 81%.In addition,user study was conducted on the calligraphy works written by the robotic arm.The questionnaire result shows that 87% of users endorse the effectiveness of beautification for handwriting and intelligent writing.In other words,our system can achieve beautification in Kai style for handwriting characters.In this thesis,we mainly study the method of calligraphy beautification from the handwritten Chinese characters for robotic arm.Through four processes including handwriting stroke extraction,calligraphy beautification from handwriting,intelligent writing and system evaluation,calligraphy works written by robotic arm can be effectively beautified to Kai style while maintaining the characteristics of users' handwriting.The method can provide a theoretical basis for the research on robot intelligent writing,and the development of intelligent calligraphy beautification and writing system for robotic arm also provide convenience for calligraphy lovers to experience the charm of Chinese calligraphy and carry forward Chinese traditional culture.